Nov 22 22:01:10 <Pupeno> Welcome everybody to the first meeting on KDE Accessibility.
Nov 22 22:01:25 <Pupeno> You can find the agenda of the meeting in
Nov 22 22:02:00 <Pupeno> and as the agenda says, we'd like you to introduce yourself (nothing long, nothing fancy, nothing you don't want to say)
Nov 22 22:02:35 <Pupeno> real name, e-mail, work on progress, e-mail address, instant messanger contacts, time constraints, comments.
Nov 22 22:03:04 <bill_H> shall I start (alphabetical) after Pupeno?
Nov 22 22:03:16 <Pupeno> There's no closing time with the meeting, if anybody wants to change the priorities in the agenda for any reason, or add new topics, feel free to just say it.
Nov 22 22:03:21 <Pupeno> bill_H: you can start.
Nov 22 22:03:51 <bill_H> sure, if you like.  I'm Bill Haneman,
Nov 22 22:04:10 <bill_H>, architect of "GNOME Accessibility Project"
Nov 22 22:04:30 <bill_H> which we voted to rename "GNU Accessibility Project" if that's not redundant,
Nov 22 22:04:50 <bill_H> but we haven't gotten FSF official approval yet, whatever that might entail :-)
Nov 22 22:05:20 <bill_H> I'm usually on gnome IRC or #gnome #sun
Nov 22 22:05:55 <bill_H> I maintain at-spi, gnome-mag, atk (co-maintain), am involved in gnome-speech, 
Nov 22 22:06:11 <bill_H> gnome-mag, 'gnopernicus' screenreader and 'gok' onscreen keyboard.
Nov 22 22:06:28 <bill_H> GAP goal is completely accessible *nix desktop
Nov 22 22:07:00 <bill_H> as a free alternative to MS
Nov 22 22:07:43 <bill_H> We have and Mozilla and Java integration already (in development, nearing alpha)
Nov 22 22:08:06 <bill_H> . that's probably enough from me ;-)
Nov 22 22:08:11 <Pupeno> ok.
Nov 22 22:08:23 <bill_H> oh - except
Nov 22 22:08:46 <bill_H> I'm here because we want to work as well together with KDE as possible
Nov 22 22:08:58 --> minux (~minux@ has joined #kde-accessibility
Nov 22 22:09:02 <bill_H> there's plenty of work for everyone, no need to do it twice ;-)
Nov 22 22:09:05 <minux> hello again
Nov 22 22:09:14 <Pupeno> welcome back minux
Nov 22 22:09:15 <danimo> bill_H: indeed...
Nov 22 22:09:30 * minux has to watch Hamilton
Nov 22 22:10:12 <Pupeno> well, next ?
Nov 22 22:10:39 <jpsc> alphabetical?
Nov 22 22:10:55 <Pupeno> jpsc: ok.
Nov 22 22:11:09 <Pupeno> whaterver, anyone.
Nov 22 22:11:21 <jpsc> I'm JP Schnapper-Casteras
Nov 22 22:11:23 <jpsc>
Nov 22 22:11:55 <jpsc> I wear a bunch of different hats -- I started the KDE Accessibility Project, the Linux Accessibility Resource Site,
Nov 22 22:12:21 <jpsc> I organize the Linux Accessibility Conferences (the 3rd one is upcoming at CSUN 2003)
Nov 22 22:12:57 <jpsc> I'm also involved in FDAWG
Nov 22 22:13:28 <minux> FDAWG ?
Nov 22 22:13:45 <jpsc> Free Desktop Accessibility Working Group --
Nov 22 22:13:59 <jpsc> comments: welcome, thanks for being here, and amen to bill_H's points about working together and there being a lot of work
Nov 22 22:14:09 <jpsc> that's all
Nov 22 22:14:34 <minux> me ?
Nov 22 22:14:40 <jpsc> sure
Nov 22 22:14:40 <Pupeno> minux: ok.
Nov 22 22:14:58 <minux> I'm Per Edin
Nov 22 22:15:04 <minux>
Nov 22 22:15:39 <minux> started to use Linux and KDE back in 98
Nov 22 22:16:26 <minux> i like accessibility but my c++ skills has only grown the last year so i've
Nov 22 22:16:41 <minux> had some ideas popup without being coded
Nov 22 22:22:53 <Pupeno> (please, when doing this, tell when you finish, so others go on).
Nov 22 22:22:57 <minux> i like my sister, :-), i started the KStorm project and go to school each day :)
Nov 22 22:18:08 * minux is finished
Nov 22 22:18:20 <Pupeno> ok, thanks.
Nov 22 22:18:32 <Gunnar> I'm Gunnar Schmi Dt.
Nov 22 22:18:44 <Gunnar>
Nov 22 22:19:00 <Gunnar> I am the author of KMouth.
Nov 22 22:20:07 <Gunnar> Currently I am studying computer science in Paderborn (Germany). In 2000/2001 I have studied two terms abroad in Ottawa (Canada).
Nov 22 22:20:31 <Gunnar> (I think that's enough from me.)
Nov 22 22:20:40 <falonaj> I'm Olaf Jan Schmidt, Gunnars twin brother
Nov 22 22:20:46 <falonaj>
Nov 22 22:21:25 <falonaj> I once thought about studying computer science, but then found out that theology is better suited for me
Nov 22 22:21:54 <falonaj> it is now a few years since i last coded something
Nov 22 22:22:30 <falonaj> but i will contribute to kde-accessibility by co-maintaining the web pages and by writing news stories
Nov 22 22:22:55 * falonaj has said enough
Nov 22 22:23:15 <Pupeno> I'm Pupeno (my official name is Josť Pablo Ezequiel FernŠndez, but I don't use it a lot, my real name is Pupeno).
Nov 22 22:23:45 <Pupeno> If someone wants to discuss the problem official vs real, and the reality itself, welcome!
Nov 22 22:23:53 <Pupeno> my e-mail address is
Nov 22 22:24:22 <Pupeno> I'm the author of Proklam but I'm new to accessibility and kinda new to KDE too (never do much development).
Nov 22 22:24:45 <Pupeno> I also did a file monitoring app for samba (never worked good) and a musical training application (unfinished yet).
Nov 22 22:25:02 <Pupeno> I'm currently unemployed and not studing officially.
Nov 22 22:25:06 <Pupeno> That's it.
Nov 22 22:25:21 <danimo> My name is Daniel Molkentin <>
Nov 22 22:25:21 <Pupeno> oh, if someone wants to contact me,
Nov 22 22:25:40 <Pupeno> now, that's it.
Nov 22 22:26:01 <danimo> I'm the current maintainer of KControl and a contributor to KDE's core for quite some time now
Nov 22 22:26:41 <danimo> I am mainly here to clarify KDE core (and maybe Qt) specific questions as they arise as good as I can
Nov 22 22:26:48 <jpsc> fabulous
Nov 22 22:27:01 <Pupeno> that's great! thank you danimo
Nov 22 22:28:16 <danimo> I am currently studying computer sience at the Technical University of Chemnitz
Nov 22 22:28:21 <danimo> that's all
Nov 22 22:28:49 <Pupeno> Well...
Nov 22 22:29:08 <Pupeno> if nothing else to say, as the agenda says, jpsc will do an introduction to kde accessibility.
Nov 22 22:29:24 <minux> fine
Nov 22 22:29:27 <jpsc> Ok, here i go
Nov 22 22:29:43 <falonaj> what about tim?
Nov 22 22:30:01 <danimo> tjansen: ping?
Nov 22 22:30:08 <danimo> falonaj: away as far as I can see..
Nov 22 22:30:14 <jpsc> yup
Nov 22 22:30:16 <falonaj> ok
Nov 22 22:30:24 <jpsc> Many of you probably already know what accessibility is, so I'll keep this quick
Nov 22 22:31:01 <jpsc> Someone said to me recently "accessibility is like usability for a different user group" and I don't think that's that far off
Nov 22 22:31:29 <jpsc> Now in terms of what technical mechanisms need to be in place for accessibility, it's a lot different than just making UI changes
Nov 22 22:32:23 <jpsc> To make software and desktops accessible, that is usable, for persons with disabilties, we need to programmaticaly expose input (keyboard, mouse movements) and output (what's on the screen) so we can modify them appropriately
Nov 22 22:32:42 <minux> such as on-screen keyboards ?
Nov 22 22:33:04 <jpsc> For someone visually impaired, that might entail sending particular text outputted to the screen to a Text To Speech (TTS) synthesize so it can be "spoken"
Nov 22 22:33:10 <jpsc> minux: yes
Nov 22 22:33:31 <danimo> too bad that we don't have a trolltech representative here because Qt has quite some unused features in place...
Nov 22 22:33:34 <bill_H> alternative input and output
Nov 22 22:33:46 <jpsc> For someone with limited mobility, an onscreen keyboard would make use of a programmatically exposed keyboard input to do fancy or non-standard things with keyboard input
Nov 22 22:33:59 <jpsc> bill_H: yes, "alternative input and output" <-- much more elegantly put
Nov 22 22:34:13 <jpsc> Point 2: What is KDE accessibility?
Nov 22 22:34:14 <Pupeno> danimo: my fault, I shold have send them an invitation.
Nov 22 22:34:31 * minux is up for writing an on-screen keyboard application
Nov 22 22:34:33 <bill_H> (next time maybe :-)
Nov 22 22:34:40 <jpsc> yes, perhaps next time
Nov 22 22:34:52 * bill_H can chat with minux about gok
Nov 22 22:34:53 <danimo> QEvent for example is prepared for speech input
Nov 22 22:35:06 <jpsc> KDE accessibility is making the K Desktop Environment work for users with disabilities
Nov 22 22:35:19 <jpsc> This a very large task -- not as simple as I'm describing it or as it may seem
Nov 22 22:35:31 <danimo> as far as I know, the entire Qt Toolkit is, see:
Nov 22 22:35:45 <jpsc> bill_H can attest to this and the scale of accessibility efforts in GUI desktops (i.e. GNOME)
Nov 22 22:36:01 <danimo> the problem there IIRC was that IBM never released ViaVoice for Linux
Nov 22 22:36:30 <falonaj> danimo: there are some important things missing in QAccessibleObject, but we can come to that later
Nov 22 22:36:42 <danimo> falonaj: ok
Nov 22 22:36:46 <jpsc> yes, we should definitely
Nov 22 22:37:22 <bill_H> danimo: the QT work was centered on MSAA, "Microsoft Active Accessibility"
Nov 22 22:37:40 <bill_H> danimo: there are some big challenges to extending that enough to work outside win32
Nov 22 22:37:53 <minux> danimo: can't we *ask* them if they have an unrelease version for Linux ?
Nov 22 22:38:10 <bill_H> danimo: they did (IBM) briefly, but withdrew it.
Nov 22 22:38:42 <jpsc> Point 3: Why does it matter?  I'm preaching to the chorus here -- in brief there are lots of reasons: moral, it's a challenge, legal / regulatory (e.g., section 508 in the U.S.), or because of the "cross-over" benefits of accessibilty features (e.g., being able to have KDE speak to you in your car, for mobile users, etc.)
Nov 22 22:38:48 <falonaj> Gunnar and i have already thought of a possible stragety to add this support to qt and kde, but we should make it a full point of its own
Nov 22 22:38:56 <falonaj> (later)
Nov 22 22:39:28 <jpsc> that's all for introduction to the tpic
Nov 22 22:39:29 <jpsc> topic
Nov 22 22:39:32 <danimo> minux: don't know about IBM, but I spoke once with Matthias Ettrich and he said we should simply contact him when need something
Nov 22 22:39:32 * bill_H stops cross-chatting
Nov 22 22:39:41 <danimo> or rather contact trolltech
Nov 22 22:39:55 <jpsc> I can talk a little bit about expereince with TrollTech I've had later
Nov 22 22:39:56 <minux> danimo: ok
Nov 22 22:40:01 <minux> danimo: sounds good
Nov 22 22:40:01 <bill_H> right.
Nov 22 22:40:07 <Pupeno> Acording to the agenda, the first issue to talk about is the web page.
Nov 22 22:40:16 <jpsc> yup
Nov 22 22:40:17 <Pupeno> to be located in
Nov 22 22:40:20 <minux> jpsc: to keep me up-to-date, whats in the 508 (briefly)
Nov 22 22:40:44 * bill_H smiles 'briefly' ?
Nov 22 22:40:45 <jpsc> minux it's a regulation that mandates U.S. federal agencies have accessible IT
Nov 22 22:40:57 <falonaj> half an hour ago i uploaded the old version into cvs, it should be shown on the server at the end of this meeting.
Nov 22 22:41:10 <jpsc> falonaj: great
Nov 22 22:41:11 <falonaj> i am working on an updated version
Nov 22 22:41:15 <jpsc> fabulous
Nov 22 22:41:23 <Pupeno> falonaj: that's great.
Nov 22 22:41:31 <minux> jpsc: ok
Nov 22 22:41:41 <jpsc> and is the updated versions goig to have some of what was discussed on the mailing list?
Nov 22 22:41:41 <falonaj> i thought there could be 5 sections
Nov 22 22:41:49 <falonaj> exactly.
Nov 22 22:42:09 <danimo> the same is planned in the EU btw
Nov 22 22:42:16 <falonaj> I'm fine with either discussion it here, or with just everyone responding to my ideas on the mailing list
Nov 22 22:42:36 <falonaj> i will just briefly describe my ideas
Nov 22 22:42:38 <Pupeno> falonaj: would you be the official maintainer of the web page ?
Nov 22 22:43:06 <bill_H>
Nov 22 22:43:36 <falonaj> Pupeno: yes
Nov 22 22:43:41 <minux> interessting
Nov 22 22:43:43 <jpsc> great
Nov 22 22:44:06 <Pupeno> falonaj: ok, great; jpsc, are you ok with that ?
Nov 22 22:44:13 <jpsc> yup
Nov 22 22:44:17 * jpsc Action Item: falonaj (Olaf) will be official maintainer of the web page
Nov 22 22:44:45 <Pupeno> falonaj: anyway, you'll get a lot of help there, at least, I'll help you as much as I can.
Nov 22 22:44:48 <falonaj> bill_H: Some people work on a new layout for all kde web pages, which will respect 508
Nov 22 22:45:06 <bill_H> falonaj: excellent!
Nov 22 22:45:10 <Pupeno> falonaj: you're free to tell us your ideas about the web pages.
Nov 22 22:45:14 <falonaj> Pupeno: Thanks, I need that. Maintaining mainly means coordinating.
Nov 22 22:45:32 <falonaj> I thought about 5 sections:
Nov 22 22:45:49 <falonaj> 1. KDE accessibility features (kcontrol, etc.)
Nov 22 22:46:33 <jpsc> Sorry to interrupt, just want to clarify
Nov 22 22:46:34 <jpsc> falonaj: you said "Some people work on a new layou for all kde web pages, which will respect 508" -- who is working on this?  Do you mean we are or we should be working on this? 
Nov 22 22:47:47 <falonaj> On the kde-www list, they were discussing ideas for ta new layout, and they agreed on a team working on that.
Nov 22 22:47:55 <jpsc> great, thanks
Nov 22 22:47:58 <falonaj> they prommised to make the new templates accessible.
Nov 22 22:48:01 <danimo> jpsc: << this and some derivates are out for evaluation
Nov 22 22:48:03 <falonaj> (WAI compliant)
Nov 22 22:48:11 <jpsc> ok
Nov 22 22:48:16 <jpsc> sorry for interrupting, continue
Nov 22 22:48:59 <falonaj> I think we should definately have a look at the templates, but we do not need to make them accessible.
Nov 22 22:49:06 <falonaj> anyway, secend section:
Nov 22 22:49:32 <falonaj> 2. Accessibility Programs like KMouth, KMag, Speaker plugins, etc.
Nov 22 22:50:12 <falonaj> Here it would be fine if the authors of the programs could write one or several pages about their software
Nov 22 22:50:14 * bill_H suggestion: call them something like 'accessibility aids' or 'assistive technologies' (for the bigger ones)
Nov 22 22:50:31 <falonaj> otherwise i will write 1 page and link to the program web pages
Nov 22 22:50:53 <minux> falonaj: fine with me
Nov 22 22:51:01 <falonaj> bill_H: assistive technologies would also be in section 1.
Nov 22 22:51:08 <Pupeno> I'll do the Proklam page obviously.
Nov 22 22:51:20 <bill_H> falonaj: ?
Nov 22 22:51:26 <falonaj> should we put the two sections together?
Nov 22 22:51:44 <bill_H> I don't understand the categories then.
Nov 22 22:52:06 <minux> i'll go for the on-screen-keyboard page (probably with anyone from GOK)
Nov 22 22:52:08 <bill_H> I assumed the first one was "accessibility features" of the core KDE desktop
Nov 22 22:52:11 <falonaj> I thought of having general KDE assistice technologies in the foirst sections,
Nov 22 22:52:21 <falonaj> independent programs in a second part
Nov 22 22:52:34 <falonaj> bill_H: exactly
Nov 22 22:52:50 <bill_H> "assistive technologies" is the conventional term for the independent (i.e add-on) programs
Nov 22 22:53:08 <falonaj> bill_H: Ok, I didn't know that.
Nov 22 22:53:23 <falonaj> so at-spi would not be an assistive technology?
Nov 22 22:53:26 <bill_H> so a screen reader is an assistive technology, but themeing and keyboard-navigation are accessibility "features"
Nov 22 22:53:49 <bill_H> falonaj: no, it's a feature or rather "built-in support" for assistive technologies
Nov 22 22:53:58 <bill_H> gok and gnopernicus would be the assistive technologies.
Nov 22 22:54:01 <falonaj> Ok. Then I understood.
Nov 22 22:54:05 <bill_H> for instance.
Nov 22 22:54:23 <bill_H> ok, just terminology then.
Nov 22 22:54:27 <falonaj> Proklam would also be a built-in support
Nov 22 22:54:35 <falonaj> (section 1)
Nov 22 22:54:38 * jpsc Action Item: jpsc ("JP") should send an e-mail to the kde-www people about WAI compliance (I mean to add this a while ago)
Nov 22 22:54:48 <bill_H> It'd be worth checking jp's sites to get the preferred terms, and maybe sites like TRACE center in Wisconsin
Nov 22 22:55:09 <bill_H> falonaj: it would?
Nov 22 22:55:24 <falonaj> it WILL go there
Nov 22 22:55:27 * bill_H guesses so, it's sort of in the gray area
Nov 22 22:56:00 <falonaj> Or are there any arguments for it being in section 2?
Nov 22 22:56:29 * jpsc wonders if we should save some of this talk for the mailing list -- we already know what basic content is going to go on and who is going to do it -- it's just a matter of working out the organization
Nov 22 22:57:00 <Pupeno> so, if I understand correctly, section 1 will be for KDE support for accessibility while section 2 will be for 'third party' software on accessibility ?
Nov 22 22:57:04 <falonaj> I agree.
Nov 22 22:57:11 <jpsc> ok
Nov 22 22:57:13 <Pupeno> agree.
Nov 22 22:57:22 <jpsc> falonaj: so we'll take that offline, on list
Nov 22 22:57:48 <falonaj> yes.
Nov 22 22:57:48 * jpsc Action Item: take web site organization discussion offline, on mailing list
Nov 22 22:58:00 <jpsc> falonaj: any other web site related points to talk about?
Nov 22 22:58:21 <falonaj> no, that's all. feel free to mail me any ideas.
Nov 22 22:58:42 <jpsc> great
Nov 22 22:58:56 <falonaj> oh, one thing
Nov 22 22:59:21 <falonaj> i already mailed the team doing the new layout. they already promised WAI aompliance.
Nov 22 22:59:29 <jpsc> ok
Nov 22 22:59:36 <jpsc> so maybe I'll take that action item off :-)
Nov 22 22:59:37 <jpsc> falonaj: thanks so much for your continued work and for volunteering for this
Nov 22 23:00:04 <jpsc> Issue 2 (this should go quickly): "Maintaining"
Nov 22 23:00:09 <falonaj> but we need to check, of course. but let's also continue this toopic on the mailing list
Nov 22 23:00:18 <jpsc> falonaj: right, for sure
Nov 22 23:00:24 <Pupeno> web page topic closed ?
Nov 22 23:00:33 * jpsc Action Item: continue WAI compliance offline, on list
Nov 22 23:00:37 <jpsc> Pupeno: it doesn't have to be, sorry if I jumped the gun
Nov 22 23:00:40 <minux> Pupeno: guess so
Nov 22 23:01:10 <falonaj> I ready with web site issues, anyone any further ideas/comments?
Nov 22 23:01:38 <Pupeno> falonaj: I have more ideas, but I'll mail you with them.
Nov 22 23:01:44 <falonaj> fine
Nov 22 23:01:52 <jpsc> Pupeno: sorry, didn't mean to cut you off
Nov 22 23:02:02 <Pupeno> jpsc: no problem.
Nov 22 23:02:14 <jpsc> Pupeno: ok, really feel free to talk about them :-)
Nov 22 23:02:35 <Pupeno> no it's ok... the Agenda is long... a lot to discuss yet.
Nov 22 23:02:42 <jpsc> yup
Nov 22 23:02:43 <jpsc> :-)
Nov 22 23:02:49 <Pupeno> we could go over some topics if there's time.
Nov 22 23:02:53 <jpsc> right
Nov 22 23:03:07 <Pupeno> going to the next topic
Nov 22 23:03:11 <jpsc> ok
Nov 22 23:03:16 <Pupeno> Mantaining.
Nov 22 23:03:27 <Pupeno> the millon dollar question.
Nov 22 23:03:31 <Pupeno> Who will do it ?
Nov 22 23:03:37 <jpsc> hehe
Nov 22 23:03:52 <jpsc> Were we going to co-maintain, all 4 of us (Iain is not here)
Nov 22 23:03:54 <jpsc> ?
Nov 22 23:04:13 <Pupeno> jpsc: yes, that was the idea.
Nov 22 23:04:46 <falonaj> my role is already clear. so what about the specific tasks for the others?
Nov 22 23:05:07 <Pupeno> I'll mantain Proklam, including my source code and other contributions (plug ins).
Nov 22 23:05:14 <Pupeno> jpsc: what else is there to do ?
Nov 22 23:05:30 <jpsc> I thought I was going to be "Team Coordinator / Leader" or something
Nov 22 23:05:44 <Pupeno> jpsc: yes, that eas the idea.
Nov 22 23:05:51 <jpsc> integrating our packages into KDE proper
Nov 22 23:05:56 <minux> if you dont mind, i'd like to maintain kosk, (K On-Screen-Keyboard, though the project isn't started)
Nov 22 23:05:58 <jpsc> that's a big one
Nov 22 23:06:25 <minux> and of course write some code
Nov 22 23:06:29 <jpsc> minux: ok
Nov 22 23:06:31 * bill_H wonders why we need another free software onscreen kbd
Nov 22 23:06:31 <Pupeno> jpsc: you said that you were going to have a busy year, so, I'd like you to have as few as posible tasks, well if you want of course.
Nov 22 23:06:40 <falonaj> I'm not sure whether it is not wiser to cooperate with the Gnome On-Screen Keyboard
Nov 22 23:07:01 <Pupeno> minux: that would be great.
Nov 22 23:07:03 <falonaj> Writing kosk does not make sence before we have at-spi support in kde
Nov 22 23:07:18 <bill_H> there's certainly work to do integrating... but gok does work to a useful extent with KDE apps already.
Nov 22 23:07:44 <minux> well, i did thought to take a look at gok, but why not make the keyboard part of gok go into and make gok/kosk front-ends ?
Nov 22 23:07:58 <bill_H> well, if we had at-spi support in kde then kosk would be redundant
Nov 22 23:08:18 <Pupeno> bill_H: the question of having the same application for gnome and kde is not a new issue and it has happened a lot of times.
Nov 22 23:08:19 <Gunnar> I've already tried to compile GOK. It seems to need half of Gnome. If you have both KDE and GNOME installed no problem off course...
Nov 22 23:08:25 <bill_H> the front-end stuff is at least IMO not significant.
Nov 22 23:08:48 <falonaj> so maybe we could de this:
Nov 22 23:08:55 <falonaj> 1. Add at-spi support to kde
Nov 22 23:09:04 <falonaj> 2. write a kde frontend for gok
Nov 22 23:09:07 <Pupeno> bill_H: but as gnome try to be independent of KDE, KDE tries to be independ of gnome too.
Nov 22 23:09:07 <bill_H> Pupeno: I do not think this has historically benefitted either codebase :-)
Nov 22 23:09:13 <falonaj> (in order of priority)
Nov 22 23:09:24 <bill_H> Pupeno: I think this is a bad move with regard to accessibility.
Nov 22 23:09:30 <Pupeno> bill_H: that's why separating back-end from front-end is a good idea.
Nov 22 23:10:08 <bill_H> what I am saying is that most linux users already have both KDE and GNOME.
Nov 22 23:10:13 <falonaj> Discussion this front-end only makes sense once kde works optimally together with the back-end
Nov 22 23:10:32 <Pupeno> bill_H: saying that kosk is redundant is like saying that kmail or konqueror is redundant too.
Nov 22 23:10:33 <bill_H> making two front ends does not add value, especially as the back ends are likely to be shared.
Nov 22 23:10:35 <minux> then we also get theming capability to the kde front-end
Nov 22 23:10:36 <falonaj> working on a kosk would first mean adding at-spi support
Nov 22 23:10:57 <bill_H> minux: note that from a 508 perspective the theming engines must interoperate anyway.
Nov 22 23:11:17 <Pupeno> falonaj: about at-spi support I think Gunnar may have some things to say.
Nov 22 23:11:32 <minux> bill_H: ok
Nov 22 23:11:40 <falonaj> should we discuss this later?
Nov 22 23:11:48 <falonaj> (at-spi?)
Nov 22 23:11:49 <bill_H> sure. May I pose a basic question?
Nov 22 23:11:56 <jpsc> sure
Nov 22 23:11:56 <falonaj> sure
Nov 22 23:11:59 <Pupeno> bill_H: yeah, of course.
Nov 22 23:12:24 <bill_H> What strategy makes sense for KDE accessibility? 
Nov 22 23:12:44 <bill_H> I mean, are we aiming for full "508" style accessibility?
Nov 22 23:13:26 <falonaj> yes, that should be the aim. that's why working on the back-ends is more important now then writing more front-ends
Nov 22 23:13:30 <minux> that would be *HUGE*
Nov 22 23:13:33 <bill_H> i.e. for KDE, or are we taking a pragmatic approach?
Nov 22 23:13:44 <bill_H> it would be huge.
Nov 22 23:13:54 <bill_H> but that doesn't make it a bad goal.
Nov 22 23:13:54 <jpsc> what do you mean "a pragmatic approach?"
Nov 22 23:14:05 <falonaj> it will be huge, but we can nevertheless agree on a strategy and start
Nov 22 23:14:20 <bill_H> I don't mean to imply that going for 508 is not pragmatic
Nov 22 23:14:32 <bill_H> but I think we have limited resources for KDE accessibility
Nov 22 23:15:08 <jpsc> true
Nov 22 23:15:11 <bill_H> so, are we saying we want KDE accessibility (possibly independent of, or in parallel to, GNOME)
Nov 22 23:15:32 <bill_H> or are we saying, "how can we make our KDE applications and libraries accessible"
Nov 22 23:15:51 <bill_H> if it's the second, then spending time on a KDE-based onscreen keyboard
Nov 22 23:16:27 <minux> if i understand everything correctly ist the latter one
Nov 22 23:16:30 <bill_H> (i.e. without using gtk+, etc.) seems less important than trying to make KDE libraries and apps work better for disabled users
Nov 22 23:16:37 <jpsc> how is "How can we make our KDE apps and libs accessible" different?  You mean make just KDE apps and libs accessible without looking at 508 interoperability and OS-wide concerns?
Nov 22 23:16:50 <bill_H> jpsc: well
Nov 22 23:16:52 <Gunnar> I would say either a small bridge to the GNOME (better: GNU) Accessibility Project or some similar approach stand-alone, but wwith bridging-possibilities.
Nov 22 23:17:08 <bill_H> there are two issues. it's possible that KDE folks feel that 508 is too big to take on.
Nov 22 23:17:19 <bill_H> I would hope not, but it's a massive undertaking.
Nov 22 23:17:32 <bill_H> or, while still aiming for eventual 508-compliance,
Nov 22 23:17:44 <bill_H> we could try and figure out what to do first, 
Nov 22 23:17:48 <minux> IMO, the best would be to make the accessiblity Linux wide and let each project make its own "front-end"
Nov 22 23:17:54 <jpsc> yes, I would definitely vote for that (what bill just said)
Nov 22 23:18:03 <bill_H> in terms of providing the most value for the groups of disabled users we think we can reach first.
Nov 22 23:18:06 <Pupeno> I agree with minux.
Nov 22 23:18:16 <falonaj> minux: agreed. but kde does not support the back-end side yet.
Nov 22 23:18:34 <Pupeno> falonaj: but, what back-end is there to support ?
Nov 22 23:18:37 <bill_H> I guess my point about GOK is that the front-end is not much of the code
Nov 22 23:18:49 <jpsc> Pupeno: AT-SPI
Nov 22 23:18:51 <falonaj> Pupeno: AT-SPI
Nov 22 23:19:07 <bill_H> well, GOK uses AT-SPI, but there is also a lot of logic in GOK itself.
Nov 22 23:19:38 <falonaj> I think we should concentrate on AT-SPI support first. Doing a different front-end can be done later, once the at-spi-kde bridge has matured
Nov 22 23:19:44 <bill_H> but of course GOK was not designed for multiple GUI front ends any more than Konqueror was designed to let you plug in GTK+ or Java widgets as your GUI
Nov 22 23:19:45 <Pupeno> that's what I'm talking about, I know we have to support AT-SPI, but then, for a on-screen keyboard, there's no back-end available.
Nov 22 23:20:14 * bill_H worries that this sounds discouraging to minux
Nov 22 23:20:31 <bill_H> I think there is plenty of work for minux to do, to help make GOK work better with KDE
Nov 22 23:20:47 <bill_H> rather than rewriting GOK using Qt, for instance.
Nov 22 23:21:04 <minux> why not put GOK code into AT-SPI
Nov 22 23:21:12 <minux> bill_H: thanks for backing me up :)
Nov 22 23:21:22 <bill_H> AT-SPI is API... it's a service
Nov 22 23:21:36 <bill_H> GOK is a client of the service.
Nov 22 23:21:55 <bill_H> And GOK does all sorts of very onscreen-keyboard-specific stuff
Nov 22 23:22:03 <Pupeno> bill_H: if KDE supported AT-SPI, GOK would work with KDE with no problem, right ?
Nov 22 23:22:11 <bill_H> whereas AT-SPI is a general control/event-monitoring/etc. service
Nov 22 23:22:21 <bill_H> Pupeno: yes, absolutely.
Nov 22 23:22:33 <bill_H> Pupeno: Actually GOK only uses a subset of at-spi.
Nov 22 23:22:40 <minux> ok
Nov 22 23:22:45 <jpsc> how about this:
Nov 22 23:22:50 <jpsc> action items: 1) 4 co-maintainers, JP as coordinator, 2) "busy" policies -- let other people know as soon as you're getting or about to get busy (via mailing list); 3) aim for AT-SPI back-end compliance; 4) hook minux up with GOK people and make sure GOK works well udner KDE
Nov 22 23:22:53 <Pupeno> bill_H: then, instead of 'hacking' into GOK to make it work with KDE, it would be better to make KDE support AT-SPI, right ?
Nov 22 23:23:20 <bill_H> Pupeno: in reality, I think there's some work in both areas.
Nov 22 23:23:33 <bill_H> Pupeno: you discover new bugs in each new environment :-)
Nov 22 23:23:40 <bill_H> but yes
Nov 22 23:23:48 <Pupeno> jpsc: I agree.
Nov 22 23:23:57 <jpsc> Pupeno: ok, great
Nov 22 23:24:02 <jpsc> minux: how do you feel about all this?
Nov 22 23:24:04 <bill_H> and very much so for gnopernicus, which would stand very little chance of being made to work without AT-SPI.
Nov 22 23:24:22 <Pupeno> jpsc: except that minux may do whatever he wants, work with gok, or at-spi-kde or anything.
Nov 22 23:24:28 <falonaj> minux: are you ok with this approach? doing the back-end stuff (at-spi) first?
Nov 22 23:24:32 <minux> jpsc: that sound good to me
Nov 22 23:24:40 <Gunnar> If I understand the GAP correctly, you have one accessibility deamon and each accessible program as well as each assistive technology connects as a client?
Nov 22 23:24:40 <jpsc> ok
Nov 22 23:24:41 <minux> falonaj: sure
Nov 22 23:25:09 <jpsc> so does anyone have changes / objectios to those action item, otherwise I'll add them?
Nov 22 23:25:17 <bill_H> Gunnar: yes... client and server technically, since every event listener is technically a "server"
Nov 22 23:25:48 <jpsc> then i think we might want to try to move on
Nov 22 23:25:51 <Pupeno> jpsc: ok with me.
Nov 22 23:25:53 <minux> jpsc: go for it
Nov 22 23:25:56 <jpsc> going once, going twice,....
Nov 22 23:25:57 * jpsc Action Items: 1) 4 co-maintainers, JP as coordinator, 2) "busy" policies -- let other people know as soon as you're getting or about to get busy (via mailing list); 3) aim for AT-SPI back-end compliance in KDE; 4) hook minux up with GOK people and make sure GOK works well under KDE  
Nov 22 23:25:58 <jpsc> added
Nov 22 23:26:22 <bill_H> jpsc: did you really mean to take on #3?
Nov 22 23:26:43 <Pupeno> I think we moved to the next topic naturaly... General plan:
Nov 22 23:26:49 <Pupeno> Ongoing projects and current developers 
Nov 22 23:26:49 <Pupeno> New developers (anyone out there?) 
Nov 22 23:26:49 <Pupeno> How can we avoid duplication (--> bill_H) 
Nov 22 23:27:09 <jpsc> bill_H: "aim for" -- yes I think so
Nov 22 23:27:18 <Pupeno> well, leaving articles outside (for later if there's no objection).
Nov 22 23:27:31 <falonaj> Pupeno: fine
Nov 22 23:27:34 <jpsc> bill_H: we can discuss the details of how we start going about that (large) task under "General Plan" -- and I think some people have started thinking about that
Nov 22 23:27:46 <bill_H> jpsc: cool. later then..?
Nov 22 23:27:58 <jpsc> yup
Nov 22 23:28:18 <jpsc> Pupeno: no objections to leaving articles for later, Olaf already wrote a great e-mail about it
Nov 22 23:28:29 <jpsc> Link: 
Nov 22 23:28:43 * jpsc Note: Substantive Issue 3 skipped
Nov 22 23:28:46 <falonaj> just mail me any ideas
Nov 22 23:28:50 <Pupeno> jpsc: ok, so, let's go on with what we where about duplication of efforts, projects and all that.
Nov 22 23:29:00 <jpsc> bill_H: I guess now :-)
Nov 22 23:29:08 <jpsc> so AT-SPI is going to be a cornerstone and a big task
Nov 22 23:29:14 <jpsc> perhaps we should talk about how to get started
Nov 22 23:29:24 <falonaj> Gunnar had some ideas
Nov 22 23:29:29 <jpsc> (this falls under "4. General Plans")
Nov 22 23:29:32 <jpsc> yup, I remember that
Nov 22 23:29:35 <jpsc> Gunnar: go for it
Nov 22 23:29:36 <falonaj> jpsc: right
Nov 22 23:30:50 <Gunnar> Well, as far as I did look into the accessibility classes of QT and the DCOP-interface there are already some techniques we could use for AT-SPI-support.
Nov 22 23:31:05 <Gunnar> However, they are not ideal.
Nov 22 23:31:46 <jpsc> there was an DCOP v. MCOP debate a while ago (never finished)
Nov 22 23:31:58 <bill_H> Gunnar: I believe the Qt accessibility work was very MS-centric
Nov 22 23:32:00 <jpsc> does it make sense to do it with one of those or at a lower level?
Nov 22 23:32:03 <bill_H> jpsc: yes
Nov 22 23:32:22 <bill_H> I see a couple of basic possibilities
Nov 22 23:32:29 <Gunnar> The QT accessibility classes seem to give accessibility information to some accessibility architecture inside QT (no way to access it from within KDE)
Nov 22 23:32:45 <bill_H> there are two convenient places to plug into the accessibility API layers in GAP
Nov 22 23:32:53 <bill_H> Gunnar: mm, that is unfortunate.
Nov 22 23:33:20 <bill_H> also the microsoft APIs are incomplete unless you also have the win32 api
Nov 22 23:33:29 <bill_H> by themselves they don't give enough info.
Nov 22 23:33:43 <bill_H> so if that is an issue, the Qt accessibility stuff would need to be extended.
Nov 22 23:33:52 <Gunnar> The DCOP mechanism already provides the properties of many widgets to other applications. However, we don't get informed about changes and we need to know all widget types.
Nov 22 23:33:52 <bill_H> we had to do a similar thing for Mozilla
Nov 22 23:34:19 <bill_H> Gunnar, jpsc: my concern in the DCOP/MCOP thing was that DCOP wasn't
Nov 22 23:34:31 <bill_H> structured enough, that there wasn't enough context for the
Nov 22 23:34:32 * jpsc notes: the position of TrollTech (I cannot speak in an official capacity or for their position now) _was_ "if you build it, we might incorporate it into Qt"
Nov 22 23:34:44 <bill_H> object-oriented info we needed.  But MCOP might do OK.
Nov 22 23:35:00 <Gunnar> As far as I see it, we could write some architecture of three parts:
Nov 22 23:35:25 <Gunnar> 1) some accessibility methods in the widgets
Nov 22 23:35:45 <Gunnar> 2) some support library within the application
Nov 22 23:36:42 <Gunnar> 3) a deamon which collects the information from the applications and bridges them to the assistive technologies.
Nov 22 23:37:09 <bill_H> a daemon?
Nov 22 23:37:18 * Pupeno asks... where can I get information on at-spi ?
Nov 22 23:37:29 <bill_H> why not just load a DLL into the app to talk AT-SPI
Nov 22 23:37:34 * minux asks... the same
Nov 22 23:37:56 <bill_H>
Nov 22 23:38:06 <bill_H> but the best place is GNOME cvs
Nov 22 23:38:22 <bill_H> unfortunately the GNOME docs generator can't parse IDL very well
Nov 22 23:38:34 <bill_H> so you have to read the IDL files for the inline docs :-/
Nov 22 23:38:43 <jpsc> perhaps some possible action items: 1) Look into various appraoches Gunnar mentioned (in widgets, support library, daemon) 2) Report back with their relative advantages and disadvantages, etc. via mailing list or another IRC meeting(s); 3) Try to re-establish contact / liason with TrollTech 
Nov 22 23:38:43 <jpsc>  
Nov 22 23:38:43 <jpsc> Who would do these action items?  (I could do the TrollTech one)
Nov 22 23:38:58 <bill_H> etc and look for at-spi/idl
Nov 22 23:39:15 <bill_H> before we leave this...
Nov 22 23:39:19 <danimo> jpsc: I could pipe some mail towards matthias
Nov 22 23:39:23 <jpsc> yes, sorry, didn't mean to leave this yet
Nov 22 23:39:27 <bill_H> I mentioned a couple of places to plug into the architecture.
Nov 22 23:39:59 <bill_H> one way is to implement the ATK interface in-process for Qt, probably via some kind of LD_PRELOAD.
Nov 22 23:40:05 <Gunnar> The first part could go partly into QT, the second part merged into what DCOP already can. If we need the deamon, we might discuss.
Nov 22 23:40:06 <bill_H> the advantages are:
Nov 22 23:40:21 <bill_H> (of the ATK approach):
Nov 22 23:40:34 <bill_H> ATK is an in-process API, and thus much easier to code to;
Nov 22 23:40:34 * jpsc Action Item: for danimo: e-mail mathias about this, perhaps CC some of us
Nov 22 23:40:51 <bill_H> and you can re-use the existing bridging/CORBA/etc. "atk-bridge" from GNOME.
Nov 22 23:41:19 <bill_H> the disadvantage is that it requires glib, but ATK does not require anything else from GNOME.
Nov 22 23:41:37 <bill_H> so that means that your accessibility solution does not depend on anything but glib.
Nov 22 23:41:48 <Pupeno> bill_H: can you explain us, I'm new to this and so other people. What is ATK ? what's the diference between ATK and AT-SPI ? etc.
Nov 22 23:42:39 <bill_H> ATK is the 'in-process' API that applications see, and which provides a simple set of C interfaces for getting accessibility info from widgets, getting event notification, etc.
Nov 22 23:42:58 <bill_H> AT-SPI is the "Assistive Technology Service Provider Interface", that is
Nov 22 23:43:12 --> SadEagle ( has joined #kde-accessibility
Nov 22 23:43:19 <bill_H> it's the "out of process" API that clients can use to get info about other apps.
Nov 22 23:43:45 <Pupeno> hello SadEagle, welcome.
Nov 22 23:43:55 <bill_H> so AT-SPI is what the assistive technologies use to get their info and do their remote control/interaction with apps.
Nov 22 23:44:20 <bill_H> But of course programming to an out-of-process IPC-type API is harder.
Nov 22 23:44:28 <minux> ah, hi SadEagle
Nov 22 23:44:32 <SadEagle> Hello.. Just stopping by to see whether there is anything I might be useful with..
Nov 22 23:44:43 <bill_H> AT-SPI is "toolkit-neutral", and language independent.
Nov 22 23:45:01 <tjansen> bill_H: looked at ATK lib: wouldnt this mean that each QWidget must become also a GObject/GInterface?
Nov 22 23:45:21 <bill_H> So for instance Java applications have their own accessibility API within the VM, but we "bridge" to the AT-SPI which is the common language.
Nov 22 23:45:29 <bill_H> tjansen: not necessarily
Nov 22 23:46:11 <bill_H> tjansen: but it means that the accessibility code for Qt (probably DLL or LD-PRELOAD) would create ATK object peers for most widgets.
Nov 22 23:46:34 <bill_H> tjansen: AtkObject derives from GObject, yes, thus the glib dependency.
Nov 22 23:46:46 <bill_H> But it need not be explicit in the Qt tookit
Nov 22 23:46:56 <bill_H> that's the role of bridging code.
Nov 22 23:46:57 * SadEagle goes to look at QAccessible docs
Nov 22 23:47:05 <tjansen> bill_H: the problem that I see is that each dev who is writing a QWidget must then write a GObject  (and know glib/GObject, fiddle with pure C programming etc)
Nov 22 23:47:21 <bill_H> Any way we do it, we end up needing to bridge between APIs.
Nov 22 23:47:30 <bill_H> But we do get to choose which ones :-)
Nov 22 23:47:50 <bill_H> If we bridge to AT-SPI, then the bridging code must understand AT-SPI and
Nov 22 23:47:50 <bill_H> CORBA
Nov 22 23:48:09 <bill_H> but if we bridge to ATK in-process (and re-use the 'atk bridge' from GNOME) 
Nov 22 23:48:11 <tjansen> bill_H: which is also a bad idea, because we probably don't want CORBA in Qt
Nov 22 23:48:33 <bill_H> tjansen: remember, only the bridge needs to know both ends of the equation.
Nov 22 23:48:34 <Gunnar> I thought of writing some ATK-similar classes for QT/KDE that provide some DCOP-AT-SPI
Nov 22 23:48:40 <tjansen> (and devs w are writing QWidgets probably don't want to implement CORBA interfaces)
Nov 22 23:48:47 <bill_H> tjansen: they don't have to
Nov 22 23:48:59 <bill_H> tjansen: but somebody has to write the bridging code.
Nov 22 23:49:15 <Gunnar> However, we would need a DCOP->CORBA bridge.
Nov 22 23:49:23 <bill_H> I think there's still some confusion on these points.
Nov 22 23:49:45 <bill_H> Gunnar: a generic DCOP-CORBA bridge would not be enough, nor would it be required.
Nov 22 23:49:50 <bill_H> though it might be nice anyhow.
Nov 22 23:49:51 * Pupeno is indeed confused.
Nov 22 23:49:55 <tjansen> bill_H: Qt doesn't know DCOP. then you would have a circular dependency
Nov 22 23:50:00 <bill_H> no no
Nov 22 23:50:17 <danimo> bill_H: yupp, CORBA has been sort of a nono in KDE. That was never because we wanted to invent our "own stuff" but simply because corba is too complex.
Nov 22 23:50:20 <minux> Pupeno: i *really* agree
Nov 22 23:50:44 <bill_H> OK, guys, do you want me to explain this?
Nov 22 23:50:57 <SadEagle> What's the problem with Qt's Accessibility plugin stuff?
Nov 22 23:50:59 <Pupeno> bill_H: I would higly apreatiate it.
Nov 22 23:51:04 <falonaj> tjansen: The idea was to have some KDE code which is a DCOP interface for Qt (as is already there, it justs needs to be extended)
Nov 22 23:51:09 <bill_H> I will do my best
Nov 22 23:51:38 <bill_H> but please don't be insulted if I get too basic, just in case.
Nov 22 23:51:48 <falonaj> SadEagle: There is no way of getting notified when there are changes, or when the cursor is moved
Nov 22 23:51:58 <bill_H> .
Nov 22 23:52:06 <falonaj> SadEagle: It does not seem to be easily extendable to KDE objects
Nov 22 23:52:38 <SadEagle> falonaj: The former is easy event-filter wise, the latter.. Yeah, that's tougher..
Nov 22 23:52:43 <bill_H> consider the general case of "bridging code".
Nov 22 23:53:01 <bill_H> If you have two APIs (or even languages) that you want to adapt to one another
Nov 22 23:53:02 <danimo> SadEagle: you can also get cursor moves via event filter, no?
Nov 22 23:53:24 <SadEagle> danimo: Depending on whether the objects below have mouse-tracking
Nov 22 23:53:24 * jpsc says: let's let bill_H explain
Nov 22 23:53:50 <danimo> SadEagle: hmm, you you name some that don't?
Nov 22 23:53:59 <bill_H> .
Nov 22 23:54:13 <SadEagle> danimo: A second, let me see which ones Keramik and HC force...
Nov 22 23:54:16 <bill_H> ready?
Nov 22 23:54:27 * bill_H waits
Nov 22 23:54:29 * jpsc coughs, let's let bill_H explain
Nov 22 23:54:36 * minux waits
Nov 22 23:54:54 * jpsc translates: can we pause the other discussion threads for a couple minutes
Nov 22 23:55:05 <SadEagle> danimo: A lot are conditional on style hints
Nov 22 23:55:17 * falonaj asks bill_H to go ahead and asks the other to pause for a moment
Nov 22 23:55:33 <bill_H> I heard some concerns about introducing dependencies into Qt and/or KDE.
Nov 22 23:56:01 <bill_H> That's reasonable.
Nov 22 23:56:07 <bill_H> But I believe that there is no need to introduce any dependencies on ATK or AT-SPI into Qt proper...
Nov 22 23:56:44 <bill_H> For instance, it is likely that whatever we do for widget accessibility support should be dynamically loaded, or LD-PRELOADED, etc. optionally.
Nov 22 23:57:11 <bill_H> right, for efficienct/smaller footprint when assistive technology support is not needed.
Nov 22 23:57:26 <bill_H> So, we have the general case of bridging code. It's like translation,
Nov 22 23:57:50 <bill_H> only the bridging code needs to know about (e.g. link to or implement) both ends of the conversation.
Nov 22 23:58:34 <bill_H> So, we could write an in-process bridge that listened to a combination of DCOP and KDE widget events or whatever is appropriate, and knew KDE and Qt API,
Nov 22 23:59:08 <bill_H> and also knew (for instance) ATK or AT-SPI.  Only that bridge would have new dependencies, not the widgets themselves.
Nov 22 23:59:26 <bill_H> Of course, the people who write the bridging code need to know both, 
Nov 22 23:59:44 <bill_H> which is why having the bridge link to one thing rather than another, may be attractive.
Nov 23 00:00:17 <bill_H> In the case of ATK, it's a simple API we're talking about, and an
Nov 23 00:00:25 * Pupeno still doesn't understand the whole change, bridge to what ?
Nov 23 00:00:39 <bill_H> in-process one, and the bridge would have to link to glib and atk but nothing else from GNOME.
Nov 23 00:01:10 <bill_H> Pupeno: Bridge in the sense that one end is in DCOP/KDE-land and the other end is in ATK-land.
Nov 23 00:01:24 <tjansen> Pupeno: basically to a CORBA client that can talk to servers  (like the ones from Gnome) that visualize the output and have special input method
Nov 23 00:01:25 <bill_H> So the bridge listens to KDE events and speaks ATK events
Nov 23 00:01:44 <bill_H> or hears ATK requests and turns them into KDE/DCOP messages, or whatever.
Nov 23 00:01:51 <Pupeno> and what happens on the other end of ATK ?
Nov 23 00:01:56 <bill_H> tjansen: not necessarily CORBA
Nov 23 00:02:06 <bill_H> tjansen: so far I am talking about ATK
Nov 23 00:02:26 <bill_H> ATK uses no CORBA, it's an in-process GObject-based API.
Nov 23 00:02:36 <bill_H> That's why ATK is simpler to code to in some ways.
Nov 23 00:02:52 <bill_H> coding CORBA servers is not rocket science but it's not exactly beginner stuff either.
Nov 23 00:03:11 <tjansen> bill_H: ok, but ATK itself doesnt have the special i/o methods. So how are they implemented?
Nov 23 00:03:27 <bill_H> There is existing code, called the "ATK bridge" which bridges between ATK and AT-SPI (CORBA).
Nov 23 00:03:48 <minux> if i got it right the "bridge" transfers and translate ATK->KDE and vice. versa.
Nov 23 00:03:51 <bill_H> Now, it doesn't matter how many layers/bridges there are, as long as they are all "thin" and "light"
Nov 23 00:04:00 <bill_H> yes, it's a translator.
Nov 23 00:04:21 <bill_H> but if the two things it's translating are reasonably well matched, it can be light and efficient.
Nov 23 00:04:44 <minux> ok
Nov 23 00:04:54 <bill_H> for instance a C++ binding to a C library, or C wrappers around a C++ library, might be very light and fast.
Nov 23 00:05:29 <bill_H> So the question that started this is, we have a couple of API layers where we can easily "plug in" to the existing accessibility architecure.
Nov 23 00:05:36 <bill_H> Gunnar suggested AT-SPI, which is fine.
Nov 23 00:05:49 <bill_H> But it means that the adapter/bridge has to know CORBA
Nov 23 00:05:54 <minux> i go for AT-SPI
Nov 23 00:06:08 <bill_H> minux: because?
Nov 23 00:06:20 --> xradon ( has joined #kde-accessibility
Nov 23 00:06:40 <bill_H> for those of you familiar with CORBA:
Nov 23 00:06:58 <Pupeno> bill_H: what does ATK do ?
Nov 23 00:07:06 <bill_H> AT-SPI being CORBA-based, means that the KDE bridge would not have to link to any GNOME libraries at all, directly.
Nov 23 00:07:31 <bill_H> However it would have to implement AT-SPI as CORBA services, which is relatively a lot of complex code.
Nov 23 00:07:33 <minux> bill_H: it seems to be the most straight forward thing to use since others use it, though having to add Corba is a disadvantage (i've never used corba though=
Nov 23 00:08:05 * minux thinks it's a lot of pros and cons to consider here...
Nov 23 00:08:21 <bill_H> ATK does what AT-SPI does mostly, but "in-process", that is, only available within the actual application program and not exported/communicated to the "outside" world.
Nov 23 00:08:39 <Pupeno> bill_H: oh, I understand.
Nov 23 00:08:54 * falonaj suggests that the concrete stragety for at-spi support in KDE is discussed on the list
Nov 23 00:08:55 <bill_H> well, the Mozilla guys chose to implement ATK and re-use the ATK-to-AT-SPI bridge.
Nov 23 00:09:14 * falonaj corrects himself: or atk support
Nov 23 00:09:26 <tjansen> bill_H: yes, but they already use GTK
Nov 23 00:09:32 <bill_H> but in the case of Java, ATK couldn't be used (since everything was in the VM), and so a -java-access-bridge for GNOME was required.
Nov 23 00:09:41 <Pupeno> bill_H: so, gnome-to-gnome is done with atk, gnome-to-rest_of_the_world atk->at-spi ?
Nov 23 00:09:51 <bill_H> tjansen: only trivially, Mozilla widgets are not GTK+ widgets.
Nov 23 00:10:11 <bill_H> Pupeno: not quite; gnome-to-gnome is also AT-SPI
Nov 23 00:10:26 <bill_H> Pupeno: ATK is only "widget-set-to-bridge"
Nov 23 00:10:38 <Gunnar> How does the outside world get information from atk?
Nov 23 00:10:45 <bill_H> since for an accessibility API to be useful it eventually has to cross process boundaries.
Nov 23 00:10:55 <bill_H> ATK doesn't cross process boundaries.
Nov 23 00:11:35 <bill_H> every application that uses ATK loads (optionally, at runtime) the shared library code for the "atk-bridge" which speaks CORBA to the rest of the accessibility stuff.
Nov 23 00:11:39 <tjansen> bill_H: in your ATK proposal: assuming I implement a new QWidget, with which interface am I supposed to add usability features?
Nov 23 00:11:55 <Pupeno> bill_H: so, atk is like a gnome->at-spi bridge, right ?
Nov 23 00:11:57 <bill_H> tjansen: there are a couple of ways;
Nov 23 00:12:21 <bill_H> Pupeno: the "atk-bridge" bridges between the ATK and AT-SPI apis.
Nov 23 00:12:35 * bill_H will post a link to a diagram later
Nov 23 00:12:45 <Pupeno> ok.
Nov 23 00:12:51 <tjansen> Pupeno: yes (but it could also support interfaces beside at-spi)
Nov 23 00:13:18 <bill_H> tjansen: so you could either implement ATK stuff in your widget (which might not be the preferred way)
Nov 23 00:13:41 <bill_H> tjansen: OR you could implement some KDE-specific accessibility API that provided the same features and was organized similarly
Nov 23 00:14:04 <falonaj> that was Gunnar's idea
Nov 23 00:14:16 <bill_H> tjansen: and then you could DLL the bridging code that converted from either the KDE stuff to ATK or to AT-SPI
Nov 23 00:14:29 <tjansen> bill_H: in other words, the first step would be to extend Qt (or let TT extend Qt) with a API for QWidgets that allows them to have accessibility features.. 
Nov 23 00:14:43 <bill_H> falonaj: but Gunnar was looking at bridging to AT-SPI, I am mentioning the possibility of bridging to ATK instead.
Nov 23 00:14:51 <tjansen> bill_H: this API could then either call ATK, or implement AT-SPI, or do something completely different
Nov 23 00:15:04 <Pupeno> the advantage of using ATK(->AT-SPI) instead of directly AT-SPI is that ATK is simpler ?
Nov 23 00:15:29 <bill_H> tjansen: yes, I think Qt needs to be extended first, or else an accessibility API needs to be written if the existing Qt one doesn't work well for KDE generally.
Nov 23 00:15:45 <bill_H> Pupeno: yeah, pretty much.  Simpler and lighter-weight.
Nov 23 00:16:24 <falonaj> bill_H: Gunnar and I were mainly discussing the approcah within KDE, we did not speak about whether to use AT-SPI or ATK
Nov 23 00:16:28 <bill_H> Pupeno: If you've ever written much out-of-process code, especially CORBA, you will know what I mean ;-)
Nov 23 00:16:39 <bill_H> falonaj: right.
Nov 23 00:16:40 <Pupeno> bill_H: why lighter ? if at-spi is at the end anyway ?
Nov 23 00:17:03 <bill_H> falonaj: I just wanted to clear up the concern that this might drag new dependencies into Qt or core KDE.
Nov 23 00:17:21 <bill_H> Pupeno: lighter in terms of the new stuff that has to be written.
Nov 23 00:17:31 <Pupeno> bill_H: oh, ok.
Nov 23 00:17:48 <bill_H> Pupeno: you are right that you end up using the AT-SPI stuff in the end.
Nov 23 00:18:06 <Pupeno> ok.
Nov 23 00:18:18 <bill_H> anyhow that's for you to think about.
Nov 23 00:18:30 <bill_H> sorry for the long digression, I hope it was somehow helpful.
Nov 23 00:18:36 <Pupeno> well... jpsc... I think that you should make an action item about discussing in the list if using directly at-spi or atk in the middle.
Nov 23 00:18:43 <Pupeno> bill_H: it was very, very helpful!
Nov 23 00:18:48 <Pupeno> bill_H: thank you!
Nov 23 00:18:55 <jpsc> Pupeno: good idea
Nov 23 00:19:34 <-- xradon has quit ("Client Exiting")
Nov 23 00:19:41 <falonaj> There are still many ideas to be discussed, but I think we should leave it for the list
Nov 23 00:19:53 <bill_H>
Nov 23 00:20:05 <jpsc> some possible action items: 1) Dicuss on the list using using at-spi or atk in the middle 2) Try to re-establish contact / liason with TrollTech  
Nov 23 00:20:19 <jpsc> who will do these action items?
Nov 23 00:20:19 <Pupeno> jpsc: I go for the first one.
Nov 23 00:20:35 <Pupeno> jpsc: all of us in the mailing list...
Nov 23 00:20:36 <jpsc> obviously we're all involved in the first one, but who should guide it / take responsibility for it
Nov 23 00:20:43 <jpsc> Pupeno? you want to?
Nov 23 00:20:55 <Pupeno> jpsc: I don't know anything about Corba or at-spi.
Nov 23 00:20:57 <jpsc> danimo: so you can do number 2?
Nov 23 00:21:11 <jpsc> Pupeno: it's ok, we'll all learn, perhaps that's another action item?
Nov 23 00:21:15 <Pupeno> to be honest I don't have much hope in Qt for various reasons.
Nov 23 00:21:22 <jpsc> ?
Nov 23 00:21:24 <Pupeno> Gunnar: are you in this ?
Nov 23 00:21:47 <bill_H>
Nov 23 00:21:51 <bill_H> for the big picture
Nov 23 00:22:03 <Pupeno> Qt is a cross platform toolkit so I don't think they'll be looking forward to implement Linux specific things and if they're getting the money from Windows, less posibilities.
Nov 23 00:22:09 <danimo> jpsc: sure, do you write a summary of this session? In this case, just fwd the stuff to me and I'll take care of the rest
Nov 23 00:22:33 <Pupeno> Qt will never know about some kewl technologies of kde, like dcop/mcop that can be used.
Nov 23 00:22:33 <jpsc> danimo: great, probably I'll write it -- it'll get written up
Nov 23 00:23:00 <jpsc> Pupeno: perhaps..., I guess that's sortof beyond our scope
Nov 23 00:23:02 <SadEagle> Pupeno: They don't have to. If Qt provides a sufficiently powerfull accessibility interface through plugins, the plugin can use platform-specific stuff.
Nov 23 00:23:02 <falonaj> Pupeno: There is already a QAccessible Object for Windows in place which just needs to be extended
Nov 23 00:23:12 <bill_H> Pupeno: if we can get Qt to extend their existing (windows-centric) accessibility support a little, KDE folks can do the rest.
Nov 23 00:23:26 <jpsc> bill_H: true
Nov 23 00:23:26 <bill_H> SadEagle: exactly.
Nov 23 00:23:36 <Pupeno> ok... then... go with that.
Nov 23 00:24:19 <Pupeno> jpsc: ok... I'll mantain and cordinate the at-spi efforts (hoping to get Gunnar's help who already been working on some ideas).
Nov 23 00:24:24 <bill_H> Pupeno: we managed to convince Mozilla to let us do a similar thing :-)
Nov 23 00:24:37 <SadEagle> bill_H: Well, usable docs are needed, too :-)
Nov 23 00:24:55 <jpsc> how about this: Action Items: 1) Dicuss on the list using using at-spi or atk in the middle (Pupeno, Gunnar) 2) Try to re-establish contact / liason with TrollTech, try to get them to extend their existing Windows accessibilit support (danimo, maybe jpsc)
Nov 23 00:25:05 <jpsc> any objections / changes?
Nov 23 00:25:29 <Pupeno> no
Nov 23 00:25:35 <jpsc> ok
Nov 23 00:25:36 <danimo> no
Nov 23 00:25:38 <minux> nope
Nov 23 00:25:43 * jpsc Action Items: 1) Dicuss on the list using using at-spi or atk in the middle (Pupeno, Gunnar) 2) Try to re-establish contact / liason with TrollTech, try to get them to extend their existing Windows accessibilit support (danimo, maybe jpsc) 
Nov 23 00:25:47 <jpsc> done
Nov 23 00:26:06 <jpsc> can we say item 4 "General Plan" is done?
Nov 23 00:26:09 <jpsc> so we can wrap this up
Nov 23 00:26:42 <jpsc> hearing no objections....
Nov 23 00:26:45 <Pupeno> jpsc: I think so... we didn't discuss
Nov 23 00:26:55 <jpsc> "5. Resources"
Nov 23 00:27:02 <Pupeno> jpsc: about any application, but I belive that's enough work for now.
Nov 23 00:27:13 <jpsc> Pupeno: yes, I agree, for sure
Nov 23 00:27:31 <jpsc> What do we need?  Someone said hardware synths would be helpful and that Iain could help with that...
Nov 23 00:28:03 * jpsc crosses off Issue 4 and pats bill_H on the back for his help
Nov 23 00:28:23 <Pupeno> well... Iain told me that he could help me getting hardware synths but I told him 'not yet'.
Nov 23 00:28:28 <bill_H> small question about this one
Nov 23 00:28:35 <Pupeno> bill_H: yes ?
Nov 23 00:28:55 <bill_H> since we're targeting a free desktop... what do you think of targeting free synths first?
Nov 23 00:29:03 <Pupeno> bill_H: of course.
Nov 23 00:29:33 <bill_H> there are only two that I know of, but maybe I am missing one (counting Festival and Flite as "one")
Nov 23 00:29:41 <Pupeno> bill_H: my priorities are Festival and FreeTTS, Gunnar told me that he was going to help with mbrola (not very free) and handifax (or something like that).
Nov 23 00:29:54 <bill_H> there's "FreeTTS" which is really a Flite port but supposedly faster and
Nov 23 00:30:08 <falonaj> hadifax needs mbrola
Nov 23 00:30:08 <bill_H> higher quality.  BSD license I believe.
Nov 23 00:30:35 <bill_H> be warned tho, FreeTTS is written in Java ;-)
Nov 23 00:30:55 <Pupeno> bill_H: yeah... it scares me ;)
Nov 23 00:30:56 <bill_H> but very nice on all other counts even if you don't care for Java.
Nov 23 00:32:20 <falonaj> Pupeno: It is true that Hadifax-Mbrola is not a free tts system, but it is the only working german tts system for free (linke in bear)
Nov 23 00:32:26 <Pupeno> jpsc: did you mention something about conferences ?
Nov 23 00:32:31 <bill_H> hmm
Nov 23 00:33:00 <bill_H> more voices for festival/FreeTTS (I think they can share voice data) would be a good thing to try and get volunteers for.
Nov 23 00:33:02 <jpsc> Pupeno: yes, I organize the Linux Accessibility Conference
Nov 23 00:33:13 <bill_H> especially since we're an international group
Nov 23 00:33:34 <Pupeno> bill_H: in the Festival mailing list, there are a lot of mails about new voices.
Nov 23 00:33:37 <jpsc> so do we need any hardware resources?
Nov 23 00:34:08 <falonaj> bill_H: There are German voices for festival, but the German text->phonem converter does not compile with the current festival version
Nov 23 00:34:14 <Pupeno> bill_H: I think a lot of want to make new voices... but they later stop, because it's too complex.
Nov 23 00:34:15 --> ranto (~ranto@ has joined #kde-accessibility
Nov 23 00:34:27 <bill_H> falonaj: :-(
Nov 23 00:34:27 <minux> jpsc: for future reference i'll take your email adress
Nov 23 00:34:42 * jpsc is
Nov 23 00:34:50 <Pupeno> bill_H: maybe the Festival group is not organizing properly the efforts to make new voices (maybe we should help a little in there).
Nov 23 00:34:52 <minux> thanks
Nov 23 00:35:27 <bill_H> the existing tools for creating new voices are hard to use.
Nov 23 00:35:34 <bill_H> so maybe that's a place to start ;-)
Nov 23 00:36:08 <Pupeno> jpsc: I don't think I need hardware except for a faster computer ;)
Nov 23 00:36:12 * jpsc wonders if expanding festival voices is a little beyond the scope of KDE accessibility
Nov 23 00:36:15 <jpsc> Pupeno: hehe :-)
Nov 23 00:36:37 <jpsc> if computer speed is _really_ _really_ holding someone back, e-mail me
Nov 23 00:36:57 <Pupeno> jpsc: are you getting support from any organization, I mean, for the disabled or anything like that ?
Nov 23 00:37:43 <minux> jpsc: is that including someone on 1GHz ? :)
Nov 23 00:37:44 <jpsc> Pupeno: the Trace R&D Center, at the University of Wisconsin-Madison is kind enough to support my work and reimburse me for many of my costs (
Nov 23 00:38:01 <jpsc> minux: sorry, that's faster than me, and I'm not "really realy" held back :-)
Nov 23 00:38:11 <Pupeno> I'm on 600Hz
Nov 23 00:38:22 <jpsc> I"m on 700, you're ok :-)
Nov 23 00:38:34 <Pupeno> yes, I'm ok... I can develop.
Nov 23 00:38:46 <jpsc> So are we good on resources, can we cross that off?
Nov 23 00:38:53 <Pupeno> my main problem is not the computer.
Nov 23 00:39:05 <minux> jpsc: think so
Nov 23 00:39:08 <danimo> ok guys, I have to go, it's almost 2am here
Nov 23 00:39:20 <minux> danimo: here too
Nov 23 00:39:22 <danimo> I'll be still logging, tho
Nov 23 00:39:27 <jpsc> yeah, we should wrap this up
Nov 23 00:39:37 * jpsc crosses of "Resources:"
Nov 23 00:39:49 <minux> 4 to go
Nov 23 00:39:52 <jpsc> regarding action items, I've done "/me Action Item" everytime we had one, so we can just search those out
Nov 23 00:39:58 * jpsc crosses of 6 :-)
Nov 23 00:40:05 <minux> 3 to go
Nov 23 00:40:07 <Pupeno> just a little note about it...
Nov 23 00:40:07 <jpsc> 7. Anything else?
Nov 23 00:40:14 <minux> Pupeno: what ?
Nov 23 00:40:43 <Pupeno> my main problem as some of you may know, is immigration (to US), so if someone knows about any organization that may help me, that would be good, for the rest, I'm ok.
Nov 23 00:41:16 <Pupeno> ok, that's it.
Nov 23 00:41:27 <minux> amnesty ? (sorry if that way out of topic)
Nov 23 00:41:48 <Pupeno> minux: I'm looking for sponsorship or something like that.
Nov 23 00:41:50 <jpsc> Pupeno: thank you for bringing that up
Nov 23 00:42:11 <Pupeno> amnesty is for people that are ilegal (I'm not ilegal) and they give you a short period of time to get a job.
Nov 23 00:42:17 <Pupeno> then you're ilegal again.
Nov 23 00:42:33 * minux now learned what amnesty does
Nov 23 00:43:49 <Pupeno> but well.
Nov 23 00:44:02 <Pupeno> let's go to topic 3, Articles.
Nov 23 00:44:10 <Pupeno> ok ?
Nov 23 00:44:14 <minux> sure
Nov 23 00:44:28 <falonaj> I thought of two articles
Nov 23 00:44:36 <jpsc> Pupeno: didn't Olaf cover that in his e-mail?
Nov 23 00:44:41 <jpsc> 
Nov 23 00:44:51 <falonaj> jpsc: exactly
Nov 23 00:45:00 <ranto> Pupeno: are you from Argentina? me too
Nov 23 00:45:27 <falonaj> Pupeno: here it is also 2 am, so i would not mind discussing this on the list
Nov 23 00:45:31 <Pupeno> jpsc: yes... then... falonaj, you may coment it, ask for help... anyone else, any other article ?
Nov 23 00:45:37 <Pupeno> that's ok with me.
Nov 23 00:45:44 <Pupeno> ranto: yes, I'm from Argentina.
Nov 23 00:45:55 * danimo is away: zZz
Nov 23 00:45:58 * bill_H 's eyelids are drooping too
Nov 23 00:46:19 * minux hands bill_H two pairs of matches
Nov 23 00:46:20 * jpsc votes that we end the meeting, seeing that we can discuss everything else on the list
Nov 23 00:46:21 <jpsc> any objections?
Nov 23 00:46:24 <jpsc> LOL
Nov 23 00:46:33 <Pupeno> I agree.
Nov 23 00:46:33 <jpsc> ...
Nov 23 00:46:37 <jpsc> Alright
Nov 23 00:46:37 <bill_H> thanks everybody.
Nov 23 00:46:48 <Pupeno> thank you everybody for comming.
Nov 23 00:46:58 <bill_H> I am really busy but please feel free to email be or ping me on irc if you want to chat :-)
Nov 23 00:46:59 * falonaj happily agrees (as Gunnar already seems to have fallen asleep a long time ago...)
Nov 23 00:47:02 * minux will absolutely come next time
Nov 23 00:47:10 <jpsc> yes, thanks for your time, the great discussion, and lets let the great work continue
Nov 23 00:47:17 <jpsc> see you on the list!
Nov 23 00:47:21 --- Pupeno has changed the topic to: KDE Accessibility official channel | Meeting ended (information will be posted) | Meeting agenda:
Nov 23 00:47:33 <Pupeno> BTW, I'll be maintainer of all the IRC stuff.
Nov 23 00:47:34 * falonaj is looking forward to all the great things that might be done soon
Nov 23 00:47:40 <minux> jpsc: i'll ask you when i need help with linux conference arrangments :)
Nov 23 00:47:45 <jpsc> I'll probably take the agendga down, so don't link to it yet
Nov 23 00:48:00 <jpsc> minux: ok, e-mail me whenever
Nov 23 00:48:01 <Pupeno> jpsc: ok.
Nov 23 00:48:08 <jpsc> thanks everyone